RESEARCH REPORT

WHEN PARENTS SEPARATE: FURTHER FINDINGS FROM THE

NATIONAL LONGITUDINAL SURVEY OF CHILDREN AND YOUTH

2004-FCY-6E

When Parents Separate: Further Findings from the National Longitudinal Survey of Children and Youth

Phase 2 of a three part project commissioned by the Family, Children and Youth Section of the Department of Justice Canada on:

The Impact of Parents’ Family Transitions on Children’s Family Environment and

Economic Well-Being: A Longitudinal Assessment

Prepared by: Heather Juby, Nicole Marcil-Gratton and Céline Le Bourdais

Centre interuniversitaire d’études démographiques Institut national de la recherche scientifique / Université de Montréal

Presented to: Family, Children and Youth Section

Department of Justice Canada

The results presented are based on analyses conducted in the Quebec Inter-University Centre for Social Statistics that provides researchers with access to the detailed longitudinal survey data

collected by Statistics Canada. The opinions expressed here do not represent the views of Statistics Canada.

The views expressed in this report are those of the authors and do not necessarily represent the views of

the Department of Justice Canada.

Aussi disponible en français

This report may be reproduced, in part or in whole, and by any means, without charge or further permission from the Department of Justice Canada, provided that due diligence is exercised in ensuring the accuracy of the materials reproduced; that the Department of Justice Canada is identified as the source department; and that the reproduction is not represented as an official version of the original report. © Her Majesty the Queen in Right of Canada, represented by the Minister of Justice and Attorney General of Canada, 2005

- i -

TABLE OF CONTENTS

EXECUTIVE SUMMARY ............................................................................................................ v

INTRODUCTION .......................................................................................................................... 1

I THE CHANGING FAMILY ENVIRONMENT AT BIRTH ............................................ 5 Married, cohabiting or alone? ............................................................................................. 6

First or second family?...................................................................................................... 10

The context at birth and family transitions ....................................................................... 13

II SPLITTING UP ................................................................................................................ 17 Recent trends in separation ............................................................................................... 17

Life with a lone parent ...................................................................................................... 19

Differences by province.................................................................................................... 19

III SHARING CUSTODY..................................................................................................... 23 Are more children living in shared custody when parents separate?................................ 26

What type of sharing patterns are most commonly selected by parents who share custody? ......................................................................................... 29

Does the choice of arrangement depend on the child’s age when parents separate?........ 30

How do shared living arrangements evolve through time? .............................................. 30

From shared to sole custody—a multivariate analysis ..................................................... 31

IV PHYSICAL CUSTODY ARRANGEMENTS AND CHILD SUPPORT: NEW INFORMATION .................................................................................................... 35

New information on custody arrangements ...................................................................... 35

New information on child support .................................................................................... 41

DISCUSSION ............................................................................................................................... 49

REFERENCES ............................................................................................................................. 53

- ii -

LIST OF TABLES Table 1.1 Proportion of children aged 2-13 years in 1996-97 who have experienced at least one family transition, by age group, and according to a) the parents’ conjugal status and b) the family context at the child’s birth, NLSCY, longitudinal cohort ..................................................................................... 14

Table 3.1 Distribution of children with separated parents in 1994-95 and 1998-99, according to type of custody and living arrangements at separation, NLSCY, Cycles 1-3................................................................................................... 26

Table 3.2 Impact of given variables on the probability that children (aged 0-15 years in 1998-99) live in shared custody or with their father rather than remaining with their mother when parents separate, NLSCY, Cycle 3 ..................................... 28

Table 3.3 Impact of given variables on the probability that children (aged 0-15 years in 1998-99) in shared custody at separation are living with their mother or father in 1998-99 rather than remaining in shared custody, NLSCY, Cycle 3 ........................................................ 32

Table 4.1 Impact of given variables on the probability that children, aged 4-15 years in 1998-99, whose parents separated in the two previous years, saw the “other parent” more or less frequently than agreed, NLSCY, Cycle 3....................................................................................................................... 38

Table 4.2 Impact of given variables on the probability that children in sole custody are consulted about living arrangements by separating parents, among children aged 4-15 years in 1998-99, whose parents separated in the previous two years, NLSCY, Cycle 3........................................................................................................ 41

Table 4.3 Proportion of child support agreements that include child support payments, among children aged 4-15 years in 1998-99, whose parents separated in the two previous years, according to various characteristics, NLSCY, Cycle 3 ....... 45

Table 4.4 Impact of given variables on the probability that a child support agreement does not include child support payments, among children aged 4-15 years in 1998-99, whose parents separated in the previous two years, NLSCY, Cycle 3........................................................................................................ 46

Table 4.5 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the method and regularity of child support payments, and the proportion of payment received, NLSCY, Cycle 3 ..................... 47

- iii -

LIST OF FIGURES Figure 1.1 Type of parents’ union at child’s birth, for children aged 2-13 years

in 1996-97, NLSCY, longitudinal cohort .................................................................. 6

Figure 1.2 Distribution of children according to parents’ conjugal status at the child’s birth for various cohorts of children, NLSCY, Cycles 1 and 3, Inter-regional variations through time.............................................. 7

Figure 1.3a Distribution of children according to parents’ conjugal status at the child’s birth, by region of Canada, 1983-84 cohorts, NLSCY, Cycle 1 .............. 8

Figure 1.3b Distribution of children according to parents’ conjugal status at the child’s birth, by region of Canada, 1997-98 cohorts, NLSCY, Cycle 3................................ 8

Figure 1.4 Family context at birth, for children aged 2-13 years in 1996-97, NLSCY, longitudinal cohort ................................................................................................... 11

Figure 1.5 Distribution of children born in two-parent families according to the family context at birth, 1983-84 and 1997-98 cohorts NLSCY, Cycles 1 and 3.......................................................................................................... 12

Figure 1.6 Distribution of children, aged 2-13 years in 1996-97, born in two-parent families by family context and by type of parents’ union, NLSCY, Cycle 2, longitudinal cohort ................................................................................................... 13

Figure 2.1a Cumulative proportions of Canadian children who experienced their parents’ separation, various birth cohorts, NLSCY, Cycle 3................................................. 18

Figure 2.1b Cumulative proportions of Canadian children born to a lone parent or experiencing their parents’ separation, various birth cohorts, NLSCY, Cycle 3 ...................................................................................................... 18

Figure 2.2a Cumulative proportions of children who have experienced their parents’ separation by age 10 years, by region of Canada, 1983-85 and 1989-91 birth cohorts, NLSCY, Cycle 3 ........................................................................................ 20

Figure 2.2b Cumulative proportions of children who have experienced a lone-parent episode by age 10 years, by region of Canada, 1983-85 and 1989-91 birth cohorts, NLSCY, Cycle 3................................................................................ 21

Figure 3.1 Distribution of minor children for whom there was a custody order at divorce, according to the type of arrangement, Canada, 1970-2000 ..................................... 24

Figure 3.2 Distribution of children whose parents separated in the two years preceding Cycles 1 (1994-95) and 3 (1998-99), according to the living arrangements at separation, NLSCY........................................................................ 27

Figure 3.3 Distribution of children, aged 0-15 years in 1998-99, in shared custody at separation, according to the type of shared living arrangements, NLSCY, Cycle 1-3................................................................................................... 29

- iv -

Figure 3.4 Living arrangements by 1998-99 for children (aged 0-15 years) who were in shared custody at separation, NLSCY, Cycle 3 .................................. 31

Figure 4.1 Living arrangements and contact with other parent in 1998-99 for children, aged 4-15 years at Cycle 3, whose parents separated in the previous two years, NLSCY, Cycle 3..................................................................................... 36

Figure 4.2 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the frequency with which the child saw the “other parent” relative to the frequency agreed, NLSCY, Cycle 3 ...................................................................................................... 37

Figure 4.3 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the reason given for less frequent contact than agreed with the other parent, NLSCY, Cycle 3 ............. 39

Figure 4.4 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the child’s involvement in decisions regarding custody and contact arrangements, by age at separation, NLSCY, Cycle 3 ...................................................................................................... 40

Figure 4.5 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the existence and type of child support agreement, NLSCY, Cycle 3 ................................................................................... 42

Figure 4.6 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the reason why no agreement was made on child support payments, NLSCY, Cycle 3......................................... 43

Figure 4.7 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the regularity of support payments and the type of support agreement, NLSCY, Cycle 3.............................................. 48

Figure 4.8 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the proportion of child support payments made and the type of support agreement, NLSCY, Cycle 3.................... 48

- v -

EXECUTIVE SUMMARY

Background

This report was commissioned by the Child Support Team of the Department of Justice Canada as part of a project using data from the “Family History and Custody” section of the National Longitudinal Survey of Children and Youth (NLSCY) to explore the impact of parents’ family transitions on children’s family environment and economic well-being. This is the first of two reports exploring the way in which parents’ conjugal behaviour shapes the family life course of their children. In this report, we focus on the beginning of the child’s life course, situating the birth within the family life course of the parents, and looking at the first and most common transition experienced by children: their parents’ separation.

The report is divided into four sections. The first two set the scene, updating and expanding earlier analyses of changes in the context at birth for children born during the last two decades of the 20th century, comparing the different regions of Canada, and exploring how this evolution influences the likelihood that parents separate. The other two sections focus on the way separated parents share responsibility for their children’s physical and economic support, with particular reference to shared “physical” custody,1 and to the new information on custody and child support collected for the first time at Cycle 3.

Research Approach

With information about the same children collected at intervals of every two years, the NLSCY is a unique data source for studying Canadian children and families at the turn of the 21st century. The present research is based on data from the first three survey cycles, conducted during the winters of 1994-95, 1996-97 and 1998-99. By Cycle 3, the original longitudinal cohort numbered approximately 15,000 children aged 4-15 years; additional samples of young children were added at Cycles 2 and 3, bringing the total number of children participating in the survey at Cycle 3 to approximately 32,000. The analyses in this report draw on different samples of these children—the complete population, particular birth cohorts, or children with separated parents—depending on the topic being explored.

Highlights

• The context at birth changed dramatically during the last two decades, with almost one-third of Canadian children born outside marriage by the turn of the century. Comparing the oldest and youngest NLSCY children, born in 1983-84 and 1997-98 respectively, shows:

• The proportion of Canadian children born within marriage dropped from 85% to 69%.

1 Throughout the text, unless otherwise clearly indicated, shared custody refers to “physical” custody, with children living part of the time with each parent. No information is available at the NLSCY regarding shared “legal” custody, where parents share responsibility for important decisions about the child without necessarily sharing living arrangements.

- vi -

• More Canadians are choosing to establish their family within a common-law union; the proportion of children born to cohabiting couples rose from 9% to 22%. Although largely the result of births in Quebec, there has also been a significant increase elsewhere in Canada, with proportions doubling in Ontario and the Prairies, and almost tripling in Eastern Canada.

• Births to single mothers increased from 6% to 10%. In the Atlantic provinces, where proportions are highest, one in six (16%) babies were born outside a union by 1997-98.

• The proportion of children born within the second family established by their mother or father rose from 11% among the oldest cohorts to 18% among the youngest.

• The probability of parental separation rose throughout Canada during the 1980s, but leveled off by the early 1990s.

• The rise was not uniform and created greater variation between the regions.

• Ontario and British Columbia registered the steepest increases during the 1980s.

• Among children born in the late 1980s, the highest proportions with parents separating before their tenth birthday were found in Quebec (32%) and British Columbia (29%).

• Rising rates of separation and out-of-union births mean more children are experiencing life in a lone parent family: one-third of the oldest cohorts (1983-84) had lived in a one-parent family by the age of 15 years, whereas children born just five years (1988-89) later reached this level by their tenth birthday.

• The 1990s saw a rapid rise in joint custody orders;2 the proportion of children in shared physical custody also rose. Although shared physical custody often transforms into sole custody over time (only 40% of children in shared living arrangements at separation were still alternating between their parents’ homes in 1998-99), the fact of the change does not hinder continued long-term involvement of both parents after separation.

• Although children’s participation in the decisions about custody increases with age, many parents decide not to involve their children in decisions about custody. However, when they do, their children’s opinion counts.

• More couples separating in the late 1990s appear to reach a child support agreement within a relatively short period after they separate. However, child support payments were expected for under two-thirds (63%) of child support agreements reached for children (aged 4-15 years in 1998-99) whose parents separated in that period at the end of the 1990s.

2 Joint custody implies that children’s “legal” custody is shared, giving parents joint responsibility for important decisions, such as children’s schooling; in some cases of joint custody, parents also share their children’s physical custody.

- vii -

• When payments are part of the support agreement, they are mostly regular and for the full amount, at least within a relatively short period (up to two years) after the separation: regular payments were made for 84% of children, and more than 90% of regular payments were paid in full.

• Private formal child support agreements are associated with more reliable support payments than private informal agreements or agreements contained in a court order.

• Shared living arrangements and sole father custody are more common in Quebec than elsewhere in Canada. Shared custody is also more durable, agreements made about custody and access are more strictly adhered to, and children are more often consulted about these arrangements in Quebec.

Implications

Taking a life course perspective on children’s family experience provides new insights with important policy implications. The pathway into a lone-parent family, for example, is linked to the type of lifestyle a child is likely to have during the episode, with separated or divorced lone-mother families having a higher standard of living on average than those established by young single mothers. With separation rates stabilizing in the early 1990s and out-of-union birth rates rising, this second entry into single-parent family life is likely to assume a greater relative importance in the next decade. Important regional differences in out-of-union birth rates means that a higher proportion of lone-parent families may be in need of financial assistance in some provinces than in others.

Sharing physical custody, even for a limited period, is associated with continued long-term involvement with both parents. However, the costs and complexity of shared living arrangements make it unworkable at times. Qualitative research into the advantages and problems with shared physical custody is needed, to give parents, mediators and others a better basis from which to judge whether shared custody is appropriate in any given case.

With the vast majority of child support payments being made regularly and on time, the problem appears to be in negotiating an agreement rather than its implementation. While the 1997 Child Support Guidelines have had some success in helping separating couples share financial responsibilities for their children, the high proportion of couples with no support agreement shows there is still some way to go. Some parents do not have the means to pay; for others, the prohibitive financial costs involved when parents fail to agree may mean no child support agreement is reached. Only qualitative research can throw more light on this question.

Finally, this report has highlighted the distinctive nature of Quebec, not only with regard to common-law unions but also in relation to the way parental responsibilities are shared at separation. Is it a result of the different approach to separation and divorce in Quebec law or is it a social rather than a legal phenomenon? This is an interesting avenue for further research, not least to discover whether the differences between Quebec and the rest of Canada extend to the impact of separation on children.

- 1 -

INTRODUCTION

This is the second of three reports examining the impact of parents’ family transitions on children’s family environment and economic well-being based on longitudinal data from the National Longitudinal Survey of Children and Youth (NLSCY), a panel survey conducted jointly by Human Resources and Development Canada (HRDC) and Statistics Canada.3 The first report focused particularly on how changes in the conjugal situation of parents affect the economic circumstances within which children are raised (Juby, Le Bourdais and Marcil-Gratton, 2003). The next two reports examine in greater depth how parents’ conjugal behaviour shapes the family life course of their children. This report focuses specifically on aspects of the most common family transition for Canadian children, and the first for most of them—their parents’ separation. The third report will explore what happens next as mothers and fathers go their separate ways.

Two themes, often absent from research on family change, are central to this report. Together they guide the analyses and provide a backdrop for social policy formulation:

• Children’s family experience needs to be viewed as a “process” that evolves in response to decisions made by parents about their own conjugal and parental life.

In other words, the concept of “family life course” should replace that of “family structure,” on which most family research is based. This is due to two main reasons

1. Structure implies solidity or permanence, and leads implicitly to the idea that children spend their lives in a given family type: intact, lone-parent or stepfamily. As a result, a large body of research has looked at the impact on children of “growing up” in a lone-parent or stepfamily, without taking into account that relatively few children do in fact spend their entire childhood in one of these family types.

2. The family structure perspective groups all families of the same “type” into a single category, ignoring the events leading up to the formation of the family. This leads to an oversimplification of reality and to false stereotyping. Take “lone-mother families” as an example. Taking into account the event immediately preceding the transition to a lone-mother family reveals two very distinct lone-mother family types: one created when a single woman has a child outside a union, the other formed when a couple who have had children together decide to separate. These two types have very little in common with each other. In the first case, mothers tend to be young, single, and have only one young child; when they enter a new union, which most do, they generally marry their partner, who becomes the first and only father the children have ever lived with. In the second case, separated mothers tend to be older, and have more and older children; a new union starts more often with cohabitation than marriage, and the new partner arrives in the life of children who lived for an extended period with their biological father (Juby, Marcil-Gratton and Le Bourdais,

3 For more information on the NLSCY, please consult the Human Resources and Skills Development Canada (Applied Research Branch of Strategic Policy Section) website at: http://www.hrsdc.gc.ca/en/cs/sp/arb/contact/contact_us.shtml#nlscy (as of September 2004).

- 2 -

2001). Only by taking a “life-course” approach to children’s family experience is it possible to provide a more complete picture of the relevant aspects of a child’s family experience.

• More and earlier parental separation means that close and significant family members do not necessarily live in the same household; the analysis of children’s family experience, therefore, has to extend beyond the residential group.

Family structure almost always refers to the residential family unit. For intact families this definition is adequate, as the residential family includes a child’s parents and siblings. For the growing number of children in other family types, however, limiting the study of the family to the residential group is problematic. Children may belong to more than one residential group, for instance. Even if they do not, close family members (most often the father) who continue to play an important role in their children’s life may not reside in the same household. To study family life without including these individuals seriously constrains the understanding of a child’s family experience. NLSCY data makes it possible to identify and include these “non-residential” family members in many of the analyses conducted here.

In these reports, we attempt to provide a broader perspective on the changing family life experience of Canadian children. Instead of viewing parents as forming a unit, we look at them as two individuals whose conjugal and parental life courses meet for a period of time, during which a child is conceived. The child’s family life course is then dependent on whether or not parents continue to follow the same path, or whether they decide to go their separate ways—an event that occurs for some children even before their birth. Subsequently, the child’s family environment is subject to decisions made separately by each parent—decisions to form or dissolve new unions, to have additional children in these new unions etc. Appreciating this diversity is an essential step in designing programs and policies to help children and families adapt to family transitions in a healthy and positive way.

What is a family transition?

Family transitions are the building blocks of the family life course. In this study, a family transition is considered to occur when parents change their conjugal status (i.e. they create or dissolve a marriage or common-law union). In addition, certain “parental status” changes are also considered as transitions: when the mother or father becomes a “stepparent” at the formation of a new union with a partner who has children from an earlier union; or when they have a child with a new partner, providing their children with a half-sibling. Some of these changes have a direct impact on the child’s residential family group, triggering a transition from one family type to another: from intact to lone-parent family when parents separate, for example. Others, such as the arrival of a new partner in the non-custodial parent’s life will not prompt a transition, but will have an impact on the child’s extended family environment.

In this report, we focus on the beginning of the life course, situating the child’s birth within the family life course of the parents, and looking at the first and most common transition experienced by children: their parents’ separation. The report is divided into four sections:

- 3 -

• Section I sets the scene, updating and expanding earlier analyses of how the context within which children born during the last two decades of the 20th century has evolved (see Marcil-Gratton and Le Bourdais, 1999). Specifically, we examine the evolution of births within or outside of a union, within marital or common-law unions, and within a first or subsequent family unit, comparing different cohorts of children and different regions of Canada. This section also looks briefly at how these factors influence children’s subsequent family life course transitions.

• Section II updates and expands information relating to parental separation, tracing the relationship between the context at birth and the likelihood that parents separate, and showing how this relationship varies through time and for different regions of Canada.

• Section III focuses on the way in which parents divide responsibilities for children when they separate, with a particular emphasis on shared custody. The different types of shared living arrangements and the evolution of these arrangements through time are considered.

• Section IV examines the new information collected for the first time at Cycle 3 concerning custody issues (e.g. whether children are consulted; whether parents keep to the visiting arrangements agreed upon) and child support arrangements (e.g. whether support agreements are adhered to; whether support agreements mean that support payments are expected) made for children when parents separate.

Family and Custody History Data

The retrospective “Family and Custody History” section of the National Longitudinal Survey of Children and Youth provides complete conjugal and parental histories of each child’s biological parents up to the time of the first survey, in the winter of 1994-95, and is brought up to date at each subsequent survey. The family history data gathered in this survey provide a new insight into the extent to which decisions parents make with regard to their conjugal and parental life affect their children’s family life course, by making it possible to reconstruct the complex family networks created by both biological parents after they separate.

However, the wealth of information about the conjugal and parental behaviour of both biological parents, on which the analyses in this section are based, brings its own problems. Linking such complex information from one cycle to another is a time-consuming process, made more so by problems of inconsistency between information reported or recorded at successive surveys. For example, several separated mothers, who were recorded in the Cycle 1 data file as not living with a new partner, reported at Cycle 2 that they were not only living with a new partner, but that this partner had also been present two years earlier (when Cycle 1 data was collected). Verifying other data revealed that, in most cases, the new partner had arrived in the household shortly before the Cycle 1 survey, and mothers probably failed to report a new relationship whose durability was still uncertain at that time. Before starting the research itself, therefore, the first step involved creating the child samples at Cycles 1 and 2, and validating the data to ensure consistency between the family information collected at the two cycles.

- 5 -

I THE CHANGING FAMILY ENVIRONMENT AT BIRTH

Changes in marital behaviour have altered radically the context into which children are born, in ways that are closely connected to how their family life course subsequently unfolds. Until recently, most children were born to married parents who were both in their first conjugal union; the only other children within their family environment at birth were full siblings, older than themselves, and living in the same household. By the end of the 20th century, however, the declining popularity and stability of marriage produced greater variation in the contexts within which children are born (Marcil-Gratton and Le Bourdais, 1999). Starting a family outside traditional marriage has become more commonplace, and rising separation and divorce rates earlier in the family life course means more men and women are having children in more than one union. The result: growing proportions of children are born outside marriage and into complex family environments.

In terms of the life course approach used here, this means that children may be born earlier or later in their parents’ life course, and to parents who may or may not have decided to marry before their birth. Some are born even before the start of their mother or father’s conjugal life course, often to young parents who have never lived with each other or with any other partner. Others, on the contrary, are born much later on, arriving into the second or third family of one or both parents. Analyses of Cycle 1 data have shown that these factors have an impact on the child’s subsequent family life course:

• One study revealed that married couples with children are more likely to stay together than those who are cohabiting, although in Quebec the difference is decreasing (Le Bourdais, Neill and Marcil-Gratton 2000).

• A second study showed a similar association between union stability and family type at birth (Juby, Marcil-Gratton and Le Bourdais, 2001). Children born into a “first family” (i.e. a family created by a couple neither of whom has children from an earlier union) are less likely to see their parents separate than are those with either maternal half-siblings living in the family, or paternal half-siblings living elsewhere. This remains true even after controlling for the fact that second families are more often created by cohabiting couples than are first families.

Clearly, the context into which a child is born is closely associated with the number and type of family transitions that occur later on during childhood. Understanding how this context has evolved in recent years, therefore, is the first step towards appreciating how complex and diverse the family life course of Canadian children is becoming. Updating and extending analyses from earlier reports (Marcil-Gratton, 1998; Marcil-Gratton and Le Bourdais, 1999), this section looks in detail at the evolution of the circumstances surrounding the birth of the NLSCY children, comparing the experience of different cohorts of children and different regions of Canada. The parts of this report that relate directly to the family life course experience of these children are based on information collected from the longitudinal sample of approximately 15,000 children included in the first two survey cycles, and aged between 2 and 13 years at Cycle 2, carried out in 1996-97. The other analyses are based on a variety of samples from the first three cycles depending on the subject under observation.

- 6 -

Married, cohabiting or alone?

Figure 1.1 shows the distribution of the longitudinal sample of children according to whether or not they were born within a union, and the legal status of the union:

• More than three-quarters (78%) of children were born to married parents, although 30% of

parents had lived together before marrying.

• Fifteen percent (15%) of children were born outside marriage, but within a union, to

common-law couples.

• The rest (7%) were born to mothers who were not living with their child’s father.

Figure 1.1 Type of parents’ union at child’s birth, for children aged 2-13 years in

1996-97, NLSCY, longitudinal cohort (N=15023)

Marriage - no prior cohabitation

48%

Not in union7%

Common-law union15%

Marriage preceded by cohabitation

30%

By including all children in the sample, however, this distribution conceals important differences between the oldest and youngest children in the survey, and between the various regions of Canada, as the next sections show.

Changes through time

The matrimonial context within which families are established evolved a great deal even during the relatively small number of years separating the oldest and youngest NLSCY cohorts. Figure 1.2 presents the distribution by context at birth for the oldest and youngest cohorts of children in the survey—those born in 1983-84 and in 1997-98.4

• The proportion of children born within marriage (preceded or not by cohabitation) dropped from 85% of the oldest children to just over two-thirds (69%) of the youngest.

4 Cohorts included for the first time at Cycle 3.

- 7 -

• The proportion of children born within a common-law union more than doubled, rising from 9% of the oldest children in the survey to 22% of the youngest.

• The proportion of births to single mothers increased from under 6% to 10%.

By the end of the 1990s, in other words, almost one-third of Canadian children were born outside marriage, although the majority were nonetheless born within a couple, to parents in a common-law union.

Figure 1.2 Distribution of children according to parents’ conjugal status at the child’s birth for various cohorts of children, NLSCY, Cycles 1 and 3, Inter-regional variations through time

60

35

25

34

9

22

610

0

10

20

30

40

50

60

70

1983-84 1997-98

Birth Cohort

%

Direct Marriage Marriage preceded by cohabitation Common-law union Not in union

Figures 1.3a and 1.3b compare five Canadian regions in terms of the evolution of the context at birth for the oldest and youngest children. Perhaps the most striking feature of the 1983-84 cohorts (Figure 1.3a) is how similar they are from one region to another. In all regions, Quebec included, at least 80% of children were born to married parents, and under one-tenth were born outside a union in all but the Atlantic provinces. A few inter-regional differences are visible, however:

• Ontario and the Prairies were the most “traditional,” in the sense that almost 90% of births occurred within marriage.

• Births to cohabiting couples were much more common in Quebec (17%) than in any other region, and out-of-union births much less so (3%).

• The highest proportions of out-of-union births were found in the provinces at the extreme west and east of Canada—in British Columbia (9%) and the Atlantic provinces (11%).

- 8 -

• These were also the only provinces where children born to a single mother outnumbered those born to cohabiting parents.

Figure 1.3a Distribution of children according to parents’ conjugal status at the child’s birth, by region of Canada, 1983-84 cohorts, NLSCY, Cycle 1 (N=3227)

5660

64

5459

28 2724

2621

7 7 7

17

996 5

3

11

0

10

20

30

40

50

60

70

British Columbia Prairie provinces Ontario Quebec Atlantic Provinces

%

Direct Marriage Marriage preceded by cohabitation Common-law union Not in union

Figure 1.3b Distribution of children according to parents’ conjugal status at the child’s birth, by region of Canada, 1997-98 cohorts, NLSCY, Cycle 3 (N=7894)

3336

46

19

26

40 39

33

26

35

16 17

11

46

23

118 10 9

16

0

10

20

30

40

50

60

70

British Columbia Prairie provinces Ontario Quebec Atlantic Provinces

%

Direct Marriage Marriage preceded by cohabitation Common-law union Not in union

- 9 -

Less than fifteen years later, the situation had changed in all regions, but towards a greater variability (Figure 1.3b). By the end of the 20th century, regional “patterns” of matrimonial behaviour were starting to emerge, with the two central regions, Ontario and Quebec, representing the extremes, at least in terms of the choice of marriage versus cohabitation as the context for starting a family. Ontario registered significantly more births to couples who married without living together first (46%) than anywhere else in the country, while Quebec recorded significantly less (19%). In fact, less than half of all births now occur within marriage in Quebec, and in some of the most predominantly francophone areas of the province the proportion has dropped to below one-quarter (Institut de la statistique du Quebec, 2000).

Moving west from Ontario, marriage is still the choice for most couples wishing to start a family—approximately three-quarters of babies in the Prairie provinces and in British Columbia are born within marriage, although more than half have parents who had lived together before marriage. East of Quebec, the situation is very different. In the Atlantic provinces,5 only three-fifths of births are marital births, while close to one-quarter of babies are born in a common-law union. This is second only to Quebec. In other words, while the rise in common-law union births in Canada is largely the result of births in Quebec, there has been a significant increase in the phenomenon in all regions, doubling in Ontario and the Prairies, and almost tripling in Eastern Canada.

The Atlantic provinces have the highest proportion of births to single mothers, with one in six (16%) babies born outside a union. This last result is unique only in its extent; throughout Canada (with the exception of British Columbia), the proportion of out-of-union births increased just as rapidly between the oldest and youngest cohorts. After years of relative stability in the rate of “out-of-union” births, this rise needs to be interpreted with caution. It does not necessarily indicate that single-motherhood is becoming more popular. It is more likely the consequence of two other demographic changes:

• Declining fertility within conjugal unions. In the current context of falling conjugal fertility, stable rates of out-of-union births result in an automatic increase in the proportion of out-of-union births.

• A changing age structure of women of reproductive age. Single mothers are usually young—in their teens or early twenties, while married women have children later on. A decline in the number of women of the age at which marital fertility is highest relative to the number of women of the age at which non-marital fertility is highest, automatically entails a rise in the proportion of out-of-union births.

During the 1970s and 1980s, the steep drop in fertility rates, which could have led to a rising proportion of out-of-union births, was compensated for by the arrival at the main childbearing ages of the baby boom generation. During the 1990s, however, the baby boom generation moved beyond the main childbearing age groups, altering the age structure of women of childbearing age. The ratio of women aged 15-24 years (highest rates of out-of-union births) to those aged

5 Except in Prince Edward Island where three-quarters of births are within legal marriage and only 6% of births are to single mothers.

- 10 -

20-34 years (highest rates of in-union births) rose from 55% in 1991, to 59% in 1996 and 66% in 2001.

First or second family?

Earlier and more frequent separation among parents in intact families means that more mothers and fathers establish a second family with a new partner. In consequence, a growing proportion of children are born into a family environment that includes older half-siblings—the children from a previous union of their mother or father. As custody is given to women in the majority of cases, children from their first family are usually living with mothers when they start a second family. This means that children born into their mother’s second family are usually born into a stepfamily—their biological father is the stepfather of their maternal half-siblings. This is less often the case when fathers start a second family. Most of the children who are born into their father’s second family live in a residential family unit that includes themselves, their biological parents and full-siblings. They are born, residentially speaking, into an “intact” family.

Earlier research has shown, however, that children born into their father’s second family have more in common with children born in stepfamilies than intact families in terms of their subsequent family life course (Juby, Marcil-Gratton and Le Bourdais, 2001). In this analysis, we consider that the “context at birth” for this group of children is very different from that of children born into their parents’ first family; they have therefore been classified separately as being born into a “quasi-intact” family. Here, we identify four types of two-parent families:

• The term “intact family,” used synonymously with “first family,” is restricted to families where neither biological parent has had children with another partner. As Figure 1.4 shows, four out of five children were born into the first family of both parents.

• The rest of the children from two-parent families were born into a “second” family, in that at least one of their parents already had children from an earlier union. Three types are considered here:

• “quasi-intact” families (5%): children’s half-siblings are living elsewhere (usually paternal half-siblings living with their mother);

• stepfather families (5%): children with maternal half-siblings only living in the family;

• stepmother or stepfather/stepmother families (3%): children with paternal half-siblings, and at times maternal half-siblings, living in the family.

- 11 -

Figure 1.4 Family context at birth, for children aged 2-13 years in 1996-97, NLSCY, longitudinal cohort (N=14949)

Intact80%

Stepmother or Stepmother/stepfather

3%

Lone parent7% Stepfather

5%

Quasi-intact -half-sibs absent

5%

Figure 1.5 illustrates the evolution of this phenomenon for children born within a conjugal union, comparing the family context at birth for the oldest (1983-84) and the youngest (1997-98) NLSCY cohorts.

• The proportion of children born within the second family established by their mother or father rose from 11% to 18%, with a corresponding drop in the proportion of births within a first intact family, from 89% to 82%.

• Among the youngest cohorts, 11% of children lived from birth in a family that included half-siblings; another 7% had half-siblings living in another household.

- 12 -

Figure 1.5 Distribution of children born in two-parent families according to the family context at birth, 1983-84 and 1997-98 cohorts NLSCY, Cycles 1 and 3

89 82

57

4 74

2

0%

20%

40%

60%

80%

100%

1983-84 1997-98

Cohorts

Stepmother or stepmother-stepfather familyStepfather familyQuasi-intact - half-siblings absentIntact

Combining these two “situation at birth” variables, Figure 1.6 shows the strong relation between the increasing popularity of common-law unions and the rise in second families. It shows:

• Couples who choose to marry without living together first are generally in their first union and rarely have children from an earlier union. As a result, more than half the children born in an intact family (56%) had parents who had married without living together first, compared with less than one-fifth of children born in other two-parent family types.

• Cohabitation is the union of choice for establishing a second family. Four-fifths of children born in a stepfamily were born to parents who had cohabited; more than half the time, parents were still cohabiting when they were born. As a result, 43% of children born in a stepfather family, and 48% in other stepfamily types, were born outside marriage.

• Quasi-intact families (with half-siblings not living in the family) are closer to stepfamilies than to intact families in terms of the type of parents’ union at birth.

- 13 -

Figure 1.6 Distribution of children, aged 2-13 years in 1996-97, born in two-parent families by family context and by type of parents’ union, NLSCY, Cycle 2, longitudinal cohort (N=13870)

56

19 18 20

32

4440 32

13

36 43 48

0%

20%

40%

60%

80%

100%

Intact Quasi-intact Stepfather Stepmother orStepmother/stepfather

Family context at birth

Direct Marriage Marriage preceded by cohabitation Common-law union The context at birth and family transitions

How are these different birth contexts associated with the way the family life course unfolds during childhood? How are they linked to the likelihood that children experience radical change in their family environment? By the second NLSCY cycle in 1996-97, almost one-quarter of children (24%) in the longitudinal sample had already experienced at least one family transition even though, at an average age of just under eight years, they were still relatively young. For the vast majority of children—those born within a conjugal union—their parents’ separation marks the first transition. For the 7% of children whose parents were already “separated” at their birth, the formation of a conjugal union is the first transition, either when their parents start living together, or when their mother or father enters a union with someone else.

Evidently, the younger children are, the less time they have had to live a family transition. As Table 1.1 shows, 18% of pre-school children had lived some change in their family situation between their birth and Cycle 2; the proportion rose to 26% among children aged 6-9 years, reaching 28% among children aged 10-13 years. Differences would be even larger were it not for the increase in separation during the 1980s which affected the younger cohorts more than the older ones.

However, age is not the only factor that raises the probability of experiencing family change. The context within which children are born also appears to have an influence on children’s family life course. Table 1.1 indicates that the proportion of children who had already lived at least one family transition by 1996-97 varied considerably according to their parents’ conjugal status:

- 14 -

• Among children whose parents had married without prior cohabitation, one child in eight (12%) had experienced a family transition.

• This proportion rose to 20% for children whose parents had cohabited before marrying.

• It more than doubled (44%) among children born in a common-law union.

• The difference between marriage and common-law unions is particularly marked in the early years. Among children aged 2-5 years in 1996-97, the proportion with separated parents was five times higher among children born to cohabiting parents than among children born to married parents with no previous cohabitation.

• Over three-quarters of children born outside a union experienced some change in their family environment. This high proportion is due to the different nature of the first transition for this group: parents who are already “separated” at their child’s birth are more likely to enter a union than couples, especially parents, are to terminate theirs.

Table 1.1 Proportion of children aged 2-13 years in 1996-97 who have experienced at

least one family transition, by age group, and according to a) the parents’ conjugal status and b) the family context at the child’s birth, NLSCY, longitudinal cohort (N=14940)

a) By conjugal status of parents at child’s birth In a conjugal union Child’s age at survey

Direct marriage

Marriage preceded by cohabitation Common-law Total

Not in a union Total

2-5 years 5.9 11.2 33.6 13.8 62.0 17.6 6-9 years 11.8 21.7 48.7 21.3 82.1 26.1 10-13 years 17.2 28.0 54.2 24.6 86.2 27.9 Total 12.3 19.9 43.6 20.0 75.7 23.9 b) By family context at birth (children born in a two-parent family)

Child’s age at survey Intact

Quasi-intact (half-siblings absent) Stepfather

Stepmother or Stepmother/ stepfather Total

2-5 years 11.7 28.4 27.8 25.2 13.8 6-9 years 17.7 43.5 40.1 43.4 21.3 10-13 years 21.2 45.7 59.0 38.5 24.6 Total 17.0 39.8 41.1 36.1 20.0

The second part of Table 1.1 confirms that the point at which children arrive in their parents’ family life course is linked to their own experience of family change: the percentage of children born in a second family experiencing their parents’ separation is much higher than among children born in a first family. Compared with children born in intact families (17%), children from quasi-intact families (whose half-siblings do not live in the household—40%), and from stepfather families (whose maternal half-siblings are present in the household—41%), are particularly likely to have lived a family transition between birth and 1996-97.

- 15 -

These figures represent the combined experience of children of different ages and from different parts of Canada. They do not indicate whether children from more recent cohorts, or from regions with high proportions of common-law union births, are more likely to experience their parents’ separation. The next section examines in greater detail the most common family transition, and the first in the lives of most children: their parents’ separation.

- 17 -

II SPLITTING UP

Children can take different pathways into their first lone-parent family episode: they may be born into it, or arrive there at some point during childhood when their parents separate or, more rarely, when a parent dies. As mentioned earlier, the pathway has a strong influence on their experience of lone-parent family life and on their subsequent family trajectory. An earlier report documented the rapid rise, during the 1970s and 1980s, in the proportion of children experiencing life in a lone-parent family and at an increasingly early age (Marcil-Gratton and Le Bourdais, 1999). Increasing rates of separation and divorce among parents establishing families during this period was the principal motor of change; data from Cycle 3 of the NLSCY makes it possible to update this information and evaluate whether this trend continued among children born in the early 1990s. It is too early, however, to assess what additional impact the growing proportion of children born to a lone parent during the late 1990s will have on the overall risk of lone-parent family life, as children from these cohorts are still too young to have experienced their parents’ separation.

For the sake of clarity, we separate the two pathways into lone-parent family life, restricting the first analysis to the most common entry, via parental separation. We then expand the analysis to include the other pathways into lone-parent family life, through the death of a parent or through a birth outside a union. In both cases, we compare the experience of children from the oldest NLSCY cohorts (born in 1983-84), with that of children born five years later (1988-89) and ten years later (1993-94), up to the age of 15, 10 and 5 years respectively. Trends are then examined for Canada as a whole and for the different regions.

Recent trends in separation

Figure 2.1a presents the cumulative proportion6 of Canadian children, born within a marital or common-law union, who experience their parents’ separation. Comparing the different cohorts shows:

• By the age of 15 years, almost 30% of children born within a couple during the early 1980s had experienced their parents’ separation, 25% of them by the age of 12 years.

• Children born only five years later, in 1988-89, reached this level (one-quarter with separated parents) three years earlier, at around the age of 9 years.

• The rapid rise in separation during the 1980s leveled off by the early 1990s, with the 1993-94 cohorts having a similar experience to that of the 1988-89 cohorts, at least during the pre-school years.

6 Calculated using the Life Table Method, a mathematical model used to describe lifelong events experienced by a cohort. A key advantage when applied to retrospective data is the possibility of including incomplete life histories of cohort members.

- 18 -

Figure 2.1a Cumulative proportions of Canadian children who experienced their parents’ separation, various birth cohorts, NLSCY, Cycle 3 (life table method)

0

5

10

15

20

25

30

35

0 2 4 6 8 10 12 14

Age of child

%

1993-94 cohorts 1988-89 cohorts 1983-84 cohorts

Figure 2.1b Cumulative proportions of Canadian children born to a lone parent or

experiencing their parents’ separation, various birth cohorts, NLSCY, Cycle 3 (life table method)

0

5

10

15

20

25

30

35

0 2 4 6 8 10 12 14

Age of child

1993-94 cohorts 1988-89 cohorts 1983-84 cohorts

- 19 -

Life with a lone parent

For children born within a union, their parents’ separation almost always marks the beginning of an episode in a lone-parent family. The growing minority of children born to a single mother also start life in this type of family. Figure 2.1b presents the cumulative percentage of children who experience a first episode in a lone-parent family, either because they were born outside a conjugal union, or following their parents’ separation, once again comparing the experience of older and younger children.

The curves are similar to those in Figure 2.1a, but start at a higher level due to the proportion of children born outside a union and therefore living in a lone-parent family from the start (age 0 years).

• Altogether, by the age of 15 years, one third of the oldest cohorts had lived in a one-parent family at some point, 30% by the time they entered their teens.

• Children born at the end of the 1980s reached similar levels approximately five years earlier (at 10 and 8 years).

• Once again, the youngest cohorts included here followed a similar path during early childhood to that taken by the 1988-89 cohorts.

Differences by province

The last section suggests a widening gap between the different regions of Canada with regard to the matrimonial context within which families are established. Do these differences extend to the durability of the family units within which children are born? Is separation more common in regions in which marriage has lost the most ground? The cumulative proportions of children whose parents separated by their tenth birthday are presented in Figure 2.2a by region of Canada. The experience of children born between 1983-85 is compared with that of those born only six years later, between 1989-1991. This comparison shows:

• Only Quebec stands out from the other provinces in the earlier cohorts (1983-85), with over one quarter of children born within a union experiencing their parents’ separation by their tenth birthday. This compares with between 16-20% in the other regions.

• The probability of separation rose throughout Canada during the 1980s. However, the rise was not uniform and created greater variation between the regions:

• Ontario registered the steepest increase, with a more than 50% rise in the probability of separation.

• The other sharp increase was in British Columbia, leaving it (at 29%), second only to Quebec (32%) in terms of how frequently, and how early on in their children’s lives, parents separate.

- 20 -

• The Atlantic and Prairie provinces were the only regions that had not caught up with Quebec’s 1983-85 levels by 1989-91. They were the only areas in which children born in a union still had a three in four chance of living with both parents at their tenth birthday.

Figure 2.2a Cumulative proportions of children who have experienced their parents’ separation by age 10 years, by region of Canada, 1983-85 and 1989-91 birth cohorts, NLSCY, Cycle 3 (life table method)

0%

10%

20%

30%

40%

British Columbia Prairies Ontario Quebec Atlantic Provinces

1983-85 cohorts 1989-91 cohorts

In addition to the proliferation of births to cohabiting couples in the 1980s, the previous section also showed growth and provincial variations in the rate of out-of-union births—the second gateway into life in a lone-parent family. Figure 2.2b shows how adding this alters the image of lone-parent family life given when parental separation alone is considered. First, including out-of-union births raises the probability of life with a lone-parent across the board. Among the 1983-85 cohorts, at least one-fifth of children had already lived with one biological parent by the age of ten years in all regions. In British Columbia, Quebec and the Atlantic provinces, this was the case for more than one-quarter of children.

- 21 -

Figure 2.2b Cumulative proportions of children who have experienced a lone-parent episode by age 10 years, by region of Canada, 1983-85 and 1989-91 birth cohorts, NLSCY, Cycle 3 (life table method)

0%

10%

20%

30%

40%

British Columbia Prairies Ontario Quebec Atlantic Provinces

1983-85 cohorts 1989-91 cohorts

Beyond this, however, it is interesting to note that the addition of out-of-union births reduces inter-regional variations considerably, as a high proportion of out-of-union births in some regions compensates for low rates of separation, and vice versa.

• Among the more recent cohorts (1989-91), three children in ten experienced life with a lone parent before their tenth birthday in all Canadian regions.

• Children born in British Columbia had the highest probability (36%), overtaking Quebec (34%).

• Despite its relatively high conjugal stability, high rates of out-of-union births meant that one-third of children born in the Atlantic provinces also experienced lone-parent family life by the age of ten years.

This analysis shows clearly that, although the vast majority of children are still born within a conjugal union, the proportion of children spending their entire childhood with both biological parents fell rapidly during the 1980s and remained at this level in the following decade. Many children share a residence with both parents for only the first part of their childhood, and the duration of this episode also declined rapidly during the period. Despite the leveling off of separation rates during the 1990s, the rise in the proportion of out-of-union births suggests that the proportion of children experiencing life in a lone-parent family will continue to rise.

- 23 -

III SHARING CUSTODY

One of the most challenging issues faced by separating parents is how to share the physical care and economic support of their children. Are the children to live with the mother or the father after separation, or will each parent create a family home and have the children with them part of the time? Who will be responsible for the financial cost of raising the children now that the household income that previously supported one household is shared between two? Reaching an agreement in the emotionally charged climate that surrounds most couples in the throes of separation is not easy. Nonetheless, many parents come to an agreement between themselves, with or without what may be costly legal involvement. At other times, they are unable to agree and the court imposes a solution based on what is perceived to be in the best interests of the child.

The divorce statistics collected by Statistics Canada, the only national data source on how parents organize care after separation before the NLSCY, suggest a trend towards greater diversity in custody arrangements. Based on this information, Figure 3.1 illustrates how custody arrangements have evolved for minor children whose custody was included in a court order from just after the 1968 Divorce Law up to the year 2000.

• Despite minor fluctuations during the 1970s and 1980s, the proportion of children put into mother, father or joint custody in 1990 was almost identical to that of 1970, with sole mother custody granted for just under three-quarters of children, and the rest divided fairly evenly between father and joint custody.

• The 1990s saw a rapid rise in the popularity of joint custody; by the year 2000, well over one-third of children (37%) were placed in the custody of both divorcing parents. Sole father custody, on the other hand, actually declined during the period, from around 15% to under 10%. These changes meant that, by the year 2000, sole mother custody was awarded for only just over half the children concerned.

- 24 -

Figure 3.1 Distribution of minor children for whom there was a custody order at divorce, according to the type of arrangement, Canada, 1970-2000

0

10

20

30

40

50

60

70

80

90

1970 1975 1980 1985 1990 1995 2000Year

%

Mother's custody Father's custody Joint custody

Source: Statistics Canada catalogues: 82-003—vol.2 no.1, Table 2 (1970-1985); vol.3 no.4 (1990); cat. 84-213 (1995, 2000)

However, divorce statistics provide an indication of legal custody rather than actual living arrangements. Although joint custody means that separated parents continue sharing all the major decisions concerning their children, it does not necessarily mean that they share their children’s physical custody. Consequently, unlike sole custody, where children generally live with the custodial parent, joint custody covers a variety of living arrangements. These statistics are also limited in terms of the information they provide on the way parents share children’s living arrangements after separation in two ways:

• First, by referring only to divorce, they provide no information on custody arrangements reached by cohabiting couples or by married couples who separate but do not divorce.

• Second, not all divorces including minor children are covered by these statistics; custody arrangements settled privately prior to divorce are not always entered on the divorce record.

By collecting information on all types of separation, and on both legal custody arrangements and actual living arrangements, the NLSCY made it possible to fill in these gaps (see Marcil-Gratton and Le Bourdais, 1999). One drawback with the Cycle 1 data used in earlier analyses, however, was that only children up to the age of 11 years were included. By the third cycle, in 1998-99, the oldest cohorts of children had reached the age of fifteen years which, coupled with samples of young children added at each cycle, made it possible to include a wider age range of children (0-15 years). In this section, we use this sample to extend the analysis of the evolution of custody arrangements through time, with particular reference to shared custody; in the rest of the text,

- 25 -

shared custody refers to shared physical custody, and is used synonymously with “shared living arrangements.”

At the NLSCY, separated parents were asked whether there was a court order concerning the child’s custody; they were then asked specifically if the court order had placed the child in shared physical custody. However, an earlier analysis of Cycle 1 data (Marcil-Gratton and Le Bourdais, 1999, Table 7) showed that this still did not mean that children had shared living arrangements, either at separation or later on. In fact, among children aged 0-11 years, under one-quarter of children with shared custody as part of the court order actually lived part of the time with each parent. More than two-thirds lived exclusively with their mother at separation, and 11% exclusively with their father. The analysis also showed that:

• Under half the children had a court order7 concerning custody arrangements.

• Shared living arrangements were more frequent when custody was settled privately.

In Table 3.1, these figures are compared with estimates made for all separations experienced by the broader sample of children, aged 0-15 years in 1998-99. Among children with custody arrangements contained in a court order, shared custody increased (up from 12.6% in 1994-95 to 17.1%), as did the proportion of these children actually sharing living arrangements (from 2.6% to 5.6%). These figures confirm, nonetheless, the tenuous relationship between shared physical custody included in a court order and actual living arrangements:

• Only one-third of children for whom shared custody was granted actually shared living arrangements between parents at separation.

• Children were considerably more likely to live in shared custody when parents came to an agreement without legal intervention (11.6%) than with it (5.6%).

7 Settled or in progress.

- 26 -

Table 3.1 Distribution of children with separated parents in 1994-95 and 1998-99, according to type of custody and living arrangements at separation, NLSCY, Cycles 1-3

Court order No court order

Custody order

Living arrangements

Living arrangements

Total

Children aged 0-11 years in 1994-95

• Mother 80.8 89.5 86.1 86.8 • Father 6.6 7.9 5.4 7.0 • Shared 12.6 2.6 8.5 6.2 Total 100.0 100.0 100.0 100.0 N 1215 1730 3276 a Children aged 0-15 years in 1998-99

• Mother 77.7 88.0 82.0 85.0 • Father 5.2 6.4 6.4 6.5 • Shared 17.1 5.6 11.6 8.5 Total 100.0 100.0 100.0 100.0 N 2060 2142 4407 b a Includes 331 children with a court order in progress at Cycle 1. b Includes 205 children with a court order in progress at Cycle 3.

Although children are still far more likely to remain in their mother’s physical care when their parents separate, this comparison suggests some movement towards a greater equality in sharing children’s physical care after separation. In the next section, we look more closely at the question of whether the trend towards increased father involvement after separation, suggested by the rise of joint custody in the divorce statistics, is also reflected in children’s living arrangements after separation. We will then explore which forms of shared living arrangements (alternating weeks, etc.) are most commonly selected by parents sharing custody, and how these arrangements evolve over time.

Are more children living in shared custody when parents separate?

The divorce statistics presented earlier indicate a rapid increase in joint custody throughout the 1990s; the information in Table 3.1 suggests that this trend extends to actual living arrangements, though to a lesser extent. Figure 3.2 compares living arrangements made at separation for children whose parents separated in the early 1990s (1992-94) with those separating in the late 1990s (1996-98); results appear to confirm this trend. In the four-year period dividing the two interviews, the proportion of children in shared living arrangements rose from 9% to 13%, while the percentage of children in their mother’s custody dropped from 85% to 78%.

- 27 -

Figure 3.2 Distribution of children whose parents separated in the two years preceding Cycles 1 (1994-95) and 3 (1998-99), according to the living arrangements at separation, NLSCY

8578

48

7 9

24

31

0%

20%

40%

60%

80%

100%

1994-95 1998-99

With father Shared - mainly fatherShared equally Shared - mainly motherWith mother

(Children aged 0-11 years) (Children aged 4-15 years)

However, with mother custody most common among younger children, and other forms of custody more common later on, it is possible that the older age of the 1998-99 sample explains this evolution. We used multinomial logistic regression techniques to evaluate how far children’s age, as well as a number of other variables, explains the apparent trend towards greater diversity of living arrangements. The model assesses the likelihood that children live in shared custody or with their father rather than residing with their mother when parents separate, and the results, in the form of odds ratios are presented in Table 3.2.8

8 Odds ratios greater than 1 indicates that the category in question increases in the likelihood of the event occurring compared with the reference category, which is always equal to 1; odds ratios below 1 decreases the likelihood. For example, in Table 3.2, children in Quebec are more than twice (2.45 times) as likely to be in shared rather than mother custody than children living elsewhere in Canada.

- 28 -

Table 3.2 Impact of given variables on the probability that children (aged 0-15 years in 1998-99) live in shared custody or with their father rather than remaining with their mother when parents separate, NLSCY, Cycle 3 (Multinomial logistic regression odds ratios1—N = 4377)

Living arrangements (with mother) Variables Shared With father

Child’s age at separation 1.143*** 1.198*** Year of separation • (Before 1991) 1.000 1.000 • 1991-1994 1.153 .858 • 1995-1998 1.487* .650* Type of agreement • (Court order, or in progress) 1.000 1.000 • No court order 1.874*** .958 Sex of child • (Boy) 1.000 1.000 • Girl .930 .861 Region of Canada • (Canada, excluding Quebec) 1.000 1.000 • Quebec 2.450*** 1.334* Reference category given in parentheses. 1 Coefficients significant at: ***p<.001; **p<.01; *p<.05.

The model confirms the great importance of children’s age at separation in the choice of living arrangements, with both shared and sole father custody significantly more likely among older children. It also confirms an increase in the popularity of shared living arrangements during the 1990s. Even controlling for age, children whose parents separated in the second half of the decade were significantly more likely to share time between the homes of both parents than those whose parents separated during the 1980s. This was all the more likely when parents reached a private custody agreement. The move towards increasing diversity in custody arrangements is particularly marked in Quebec. Children were significantly more likely to share living arrangements in Quebec than in the rest of the country. They were also more likely to live with their father after separation, going against the general movement away from sole father custody during the period. Finally, and perhaps surprisingly, the child’s sex does not appear to have a significant impact on the type of custody.

As we have seen, shared custody is an arrangement that is not easily defined, and not only because there is no direct equivalence between a shared custody agreement and shared living arrangements. Shared living arrangements are flexible and, unlike sole custody, tend to evolve into other living arrangements with the passage of time since separation (Juby, Le Bourdais & Marcil-Gratton, 2003). In the following sections, we will explore what insight the NLSCY can give into the way parents who decide on shared physical custody of their children actually arrange this care, and how these arrangements evolve through time.

- 29 -

What types of sharing patterns are most commonly selected by parents who share custody?

For parents who decide to share the physical care of their children at separation many choices are open. The most popular options selected by parents9 at the time of separation are shown in Figure 3.3. The fact that close to a quarter came to an arrangement other than those offered in the questionnaire indicates that there are many more possibilities open.

Figure 3.3 Distribution of children, aged 0-15 years in 1998-99, in shared custody at separation, according to the type of shared living arrangements, NLSCY, Cycle 1-3 (N=373)

23

17

18

5

8

30

0 10 20 30 40

Other schedule

One weekend out of two

Weekdays/weekends

Alternate nights

Alternate 2 weeks

Alternate weeks

%

• Alternating each week between the parents is the most common form of sharing, with three

children in ten in shared custody moving from their mother’s to their father’s home each week.

• Much less commonly chosen at separation (8%) is a similar arrangement, but with children changing residence every two weeks.

• Spending alternate nights with each parent was chosen even more rarely (5%), and is probably only workable when parents live in very close proximity to one other.

• Almost one child in five (18%) spent the weekend with one parent, weekdays with the other. This may be the most practical option when one parent lives closer to the school than the other.

9 By Cycle 3, the parents of approximately 400 children organised to share their children’s physical custody when they separated.

- 30 -

• Interestingly, 17% of children declared living in shared custody spent only one weekend in two with one of their parents, and the rest of the time with the other. Time-sharing here is very similar to that arranged for many children considered to be in the sole custody of their mother or father.

Does the choice of arrangement depend on the child’s age when parents separate?

Although all types of arrangements are found at all ages, there appear to be some age-related patterns.10 The most common arrangement, alternating weeks, was chosen most frequently when children were aged between 3 and 11 years when their parents separated, less often for very young children, and rarely for adolescents. Interestingly, among adolescents, only one of the “standard” arrangements was selected with any frequency: spending weekends with one parent, weekdays with the other. Most of the others have an “other” arrangement, perhaps reflecting the greater autonomy of this age group both in terms of designing and implementing more diverse and flexible types of sharing. The standard sharing patterns are also less common among very young children. In fact, approximately 30% of children under the age of three in shared living arrangements lived primarily in one household, spending only one weekend in two with the other parent.

How do shared living arrangements evolve through time?

Only 40% of children who were in shared custody at separation were still alternating between their parents’ homes by Cycle 3 and, of these, one quarter had a different arrangement from that first put in place (Figure 3.4). Half the children were living full-time with only one parent, 32% with their mother and 17% with their father. The rest (11%) were living with both parents in the same household after the latter had decided to get back together again at some point during the period. Two types of shared living arrangements were particularly likely to transform into sole custody:

• Alternating every two weeks. Four-fifths of children with this arrangement were living in the sole custody of their mother or father, in equal proportions, by 1998/99. It is possible that this arrangement was selected at the start by parents whose circumstances were not particularly favourable to sharing custody.

• One weekend in two. As might be expected, the majority of children were declared living full time with their mother by 1998/99. It is possible that the frequency of contact with their father did not change during the period; the change in declaration may simply reflect a change in the perception of the respondent.

10 This data is not presented due to problems of confidentiality with many of the categories.

- 31 -

Figure 3.4 Living arrangements by 1998-99 for children (aged 0-15 years) who were in shared custody at separation, NLSCY, Cycle 3 (N=318)

Shared custody - other arrangement

10%

Living with father17%

Parents reunited11%

Shared custody - no change

30%

Living with mother32%

The move from shared custody to another form occurs gradually over time. Among children whose parents had been separated for under two years in 1998-99, 85% were still in some kind of shared custody arrangement. This proportion dropped rapidly with the time elapsed since separation—to 37% for those whose parents had been separated for 2-3 years, and to 13% after 4-5 years. Among children whose parents had been separated six years or more, only 8% were still living in two households.

From shared to sole custody—a multivariate analysis

The importance of the duration since separation comes out clearly in the multivariate analysis presented in Table 3.3, which evaluates the impact of various factors on the likelihood that children moved into sole mother or sole father custody by 1998-99 rather than remaining in shared custody. This analysis shows:

• The longer the time since separation, the more likely the move into sole mother or sole father custody becomes.

• Shared living arrangements put in place for children whose parents separate when they are of secondary-school age are significantly less durable than those made for primary-school aged children. This result is at first surprising. One might assume that the greater independence of older children would facilitate the movement between parents’ households, making them more likely to remain in shared custody. This does not appear to be the case. Having spent a larger part of childhood with both parents in a single residence, are they less willing to put up with the inconvenience of two homes? Does their greater autonomy in fact facilitate frequent contact with both parents, whether or not they actually live with them?

- 32 -

• Shared custody is also more likely to develop into sole father custody when children are very young at separation. This result was also unexpected, given the greater propensity towards mother custody among young children. Could it be that, at times, shared living arrangements are chosen for young children by couples in which the father took the main responsibility for the children because the mother found it hard to cope, and is unable to cope alone after separation? More detailed, qualitative data are necessary to understand this process, as they are to understand the circumstances surrounding father custody more generally.

• Girls are significantly less likely than boys to move from shared custody to father custody.

• Certain types of shared custody arrangements also influence the move to father custody: children who alternated between parents every two weeks, or who were in some other type of shared arrangements, are more likely to be with their father later on.

• Children living in Quebec in 1998-99 were significantly less likely to move from shared living arrangements to living exclusively with either their mother or father. Not only is shared custody more common in Quebec than elsewhere in Canada (see Table 3.2), it is also more durable.

Table 3.3 Impact of given variables on the probability that children (aged 0-15 years in 1998-99) in shared custody at separation are living with their mother or father in 1998-99 rather than remaining in shared custody, NLSCY, Cycle 3 (Multinomial logistic regression odds ratios1—N = 269)

Living arrangements at C3 (shared custody) Variables

With mother With father Duration since separation 2.145*** 2.274*** Child’s age at separation • (6-11 years) 1.000 1.000 • 0-5 years .968 3.361* • 12-15 years 4.570* 14.371*** Sex of child • (Boy) 1.000 1.000 • Girl 1.832 .162*** Custody arrangement at separation • (Alternating weeks) 1.000 1.000 • Alternating two weeks 1.077 8.194* • Alternating nights .721 • Weekdays / weekends .469 1.647 • One weekend in two 1.844 .222 • Other .606 5.067* Region of Canada • (Canada, excluding Quebec) 1.000 1.000 • Quebec .398* .264**

Reference category given in parentheses. 1 Coefficients significant at: ***p<.001; **p<.01; *p<.05.

- 33 -

Beyond this, however, the question remains—why are shared custody arrangements so likely to transform into sole custody over time? Is it too difficult to maintain as an arrangement? Where does the impetus for change come from? From the parents, because of some external change, in the place of work or new family commitments, or from the child who finds moving between two homes unsettling or complicated? Unfortunately, this sort of question cannot be answered by the NLSCY, which provides little information on the circumstances of the family unit that does not contain the principal respondent for the child. Even the distance separating the homes of separated parents is unknown. Nonetheless, one important fact can be gleaned from the data: however shared custody may evolve in the years following separation, it appears to be associated with a higher level of children’s continued involvement with both parents after separation. Even when mothers and fathers no longer shared the physical custody of their children, over 60% of the “non-resident” parents maintained at least weekly contact with their children; more than three-quarters saw them at least every two weeks, and fewer than one in twenty had not seen their child in the twelve months preceding Cycle 3. This frequency of contact is considerably higher than that found between non-resident parents and children living in sole custody from the moment of separation.

- 35 -

IV PHYSICAL CUSTODY ARRANGEMENTS AND CHILD SUPPORT: NEW INFORMATION

Only limited information on custody and child support was collected at the first two NLSCY cycles. An attempt was made to introduce more detailed questions in subsequent cycles. Unfortunately, problems with the complex questionnaire meant that this information was not collected exhaustively and, as a result, the analyses possible from the data are not as broad as had been hoped. The information is complete, however, for all parental separations occurring between Cycles 2 and 3 among children in the longitudinal cohort, and the analyses in this section are based on these children. The following analyses refer, therefore, to separations occurring during the two years preceding Cycle 3, among parents of children aged 4-15 years in 1998/99 (i.e. to children aged between 2 and 15 years when their parents separated.) Using only recent separations has certain advantages. First, it improves reliability, as the information about decisions taken at separation is still fresh in the respondent’s mind. Second, the information is “up-to-date,” describing the situation right at the end of the 1990s.

New information on custody arrangements

The distribution in Figure 4.1 provides a general picture of the custody and visiting arrangements in place by 1998-99 for children whose parents had separated in the two previous years. Most parents had maintained close ties with their children within the relatively short period of time since the separation:

• The parents of 13% of the children were living together again by 1998-99.

• Almost two-thirds (42%+14%+7%) were living with their mother; of these, two-thirds saw their father on a regular weekly or two-weekly basis.

• Only one child in fourteen (7%) was living with their father full-time; of these, just under half saw their mother on a regular weekly or two-weekly basis.

• One child in eight (12%) was living in shared custody, alternating between their mother’s and father’s home.

• A minority of children (5%) fell outside the standard categories for a variety of reasons (e.g. in the care of someone other than a biological parent; a parent had died; a different form of contact).

- 36 -

Figure 4.1 Living arrangements and contact with other parent in 1998-99 for children, aged 4-15 years at Cycle 3, whose parents separated in the previous two years, NLSCY, Cycle 3 (N=867)

Other5%

With father, sees mother irregularly

4%

Parents reunited13%

With mother, never visits father7%

With mother, sees father irregularly

14%

Shared living arrangements12%

With father, sees mother regularly

3%

With mother, sees father regularly

42%

In other words, 70% of the children were still in close contact with both parents at Cycle 3; under one-fifth (14%+4%) had irregular contact with the other parent, and only 7% of children had no contact at all. But how does this compare with the amount of time parents had actually arranged to spend with their children? For the first time at Cycle 3, the parent responding to the survey was asked about the amount of time the child had spent with the other parent during the last year (or since the separation, if it happened less than a year ago) relative to what had been agreed; results show a high degree of flexibility in arrangements.

Do parents keep to the visiting arrangements agreed on at separation?

Overall, a little under half (46%) of these children spent the amount of time agreed upon with their other parent (see Figure 4.2). Parents not keeping to the agreement were divided fairly evenly between those who saw their child much more, a little more, a little less and much less frequently than agreed. Only 2% of children had not seen the other parent in the last year (or since the separation, if it was under a year ago); another 1% had no visitation rights. Clearly, arrangements made when parents separate are not fixed, and evolve in very different ways. Why should this be so? Additional information gathered for the first time at Cycle 3 throws some light on this question.

- 37 -

Figure 4.2 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the frequency with which the child saw the “other parent” relative to the frequency agreed, NLSCY, Cycle 3 (N=856)

3

1

2

12

14

46

12

12

0 10 20 30 40 50

Other

No visitation rights

No contact

Much less often

Less often

Amount agreed

More often

Much more often

%

The multivariate analysis presented in Table 4.1 evaluates the impact of various factors on the likelihood that parents see their children more or less frequently than agreed.11 Findings from the model show that two variables in particular are linked to whether or not agreements are kept to:

• Seeing children either more frequently or less frequently than agreed is significantly less likely in shared custody arrangements than in sole custody. In other words, when parents agree to share living arrangements, the timetable is more strictly adhered to than when children live exclusively with their mother or father.

• Families living in Quebec are also more likely to keep to the arrangements made, even after controlling for the greater incidence of shared custody in the province.

11 The “less frequent contact” category includes children with no contact during the previous year; those with no visiting rights were excluded from the analysis.

- 38 -

Table 4.1 Impact of given variables on the probability that children, aged 4-15 years in 1998-99, whose parents separated in the two previous years, saw the “other parent” more or less frequently than agreed, NLSCY, Cycle 3 (Multinomial logistic regression odds ratios1—N = 807)

Variables More frequently Less frequently Living arrangements • (With mother) 1.000 1.000 • With father 0.701 0.581 • Shared 0.281*** 0.339*** Type of custody agreement • (Court order) 1.000 1.000 • No court order 0.679+ 0.958 Child support agreement • (Yes) 1.000 1.000 • No 1.159 0.993 Child’s age at separation • (2-5 years) 1.000 1.000 • 6-11 years 1.078 0.764 • 12-15 years 0.508* 0.890 Sex of child • (Boy) 1.000 1.000 • Girl 1.832 1.121 Region of Canada • (Canada, excluding Quebec) 1.000 1.000 • Quebec .304* .242*** Reference category given in parentheses. 1 Coefficients significant at: ***p<.001; **p<.01; *p<.05; +p<.1.

In addition, children who are older when parents separate are less likely to see the other parent more frequently than agreed than are younger children. This finding may have no relation to the separation itself; it may simply reflect the fact that children generally spend less time with parents and more time with their peer group as they enter their teens. More frequent contact than agreed is also less likely (at marginal significance) when custody was settled privately.

Why do parents see their children less often than agreed?

Interestingly, none of the other variables contained in the analysis are associated with less frequent contact, probably because the explanation is to be found elsewhere. A question was added at Cycle 3 aimed at providing some insight into this question: parents who stated that contact over the last year with the other parent was less frequent than agreed, or absent, were asked the main reason, or reasons, for this.

The distribution of replies is as follows (Figure 4.3):

• The responding parent cancelled the visit—4%

- 39 -

• The non-resident parent cancelled the visit—38%.

• The non-resident parent has not tried to see the child over the last year—34%.

• The child did not want to see the non-resident parent—24%.

Figure 4.3 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the reason given for less frequent contact than agreed with the other parent, NLSCY, Cycle 3 (N=259)

24

34

38

4

0 10 20 30 40

The child did not want to seethe other parent

The other parent has not triedto see the child

The other parent cancelledthe visit

The responding parentcancelled the visit

%

Respondent rarely claimed responsibility for visits being cancelled. The child’s wishes were stated as the cause for close to one-quarter of cases, but mostly the visiting parent was declared directly responsible for the lower levels of contact. Evidently, these responses need to be interpreted within their context. This question is highly subjective, particularly given the relatively short time since separation. Many parents may still have problems relating to one another (and to their children), with access to children being a particularly touchy issue, and their perception of the source of problems relating to this is likely to differ. Unfortunately, the question could not be put to the other parent, making it impossible to assess the extent of any differences between separated parents. In addition, the choice of possible responses is very general, revealing little or nothing about the dynamic lying behind the stated cause.

Do parents consult their children about custody and visiting arrangements?

Children rarely participate directly in formal divorce, custody and access proceedings in Canada. A recent report commissioned by the Department of Justice recommended that they should be given more of a voice (Bessner, 2002). Most parents settle custody privately, however. Do children have a more direct say in these arrangements? Does it depend on their age when parents separate? For the first time in 1998-99, parents were asked about their child’s involvement in the decisions made about custody and visiting arrangements. Unfortunately, the question was put

- 40 -

only for children in the sole custody of mother or father; how far children were involved in arranging shared custody arrangements remains a mystery. Overall, among children aged 4-15 years, whose parents separated in the second half of the 1990s and who were placed in sole custody:

• The majority (71%) had not been consulted, more than half (56%) because they were considered too young.

• Among those who had been consulted (29%), their wishes prevailed the majority of the time (18%); for the others, parents asked their opinion, but took the final decision themselves.

That children’s participation in the decision-making process is closely related to their age at the time of separation is clearly illustrated in Figure 4.4. The likelihood that children are involved in the discussions about their future rose dramatically with age, and the proportion of children considered too young to be consulted fell just as rapidly, with no parent declaring a teenage child too young. Nonetheless, many parents decide not to involve their children in the decision irrespective of age. Almost one-quarter of teenagers had no say in the custody arrangements made by their separating parents. On the other hand, when parents decide to consult their children on this issue, their opinion counts. Almost half the teenagers were not only consulted, they also had the final say in the decision. This was true for one-third of children aged 10-12 years, and one-fifth of those aged 7-9 years. In any case, even when parents make the final decision themselves, their children’s wishes are likely to have been taken into account.

Figure 4.4 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the child’s involvement in decisions regarding custody and contact arrangements, by age at separation, NLSCY, Cycle 3 (N=743)

8677

47

28

1114

19

28

23

7

13

12

28

2132

49

3 2

0

20

40

60

80

100

0-3 years 4-6 years 7-9 years 10-12 years 13-15 years

Child's age at separation

%

Consulted-child's wishesprevailed

Consulted-parentsdecided

Not consulted

Too young

- 41 -

What other factors have an influence on whether or not children are consulted? Are children consulted as often when custody is subject to a court order even though they rarely have a direct say in the proceedings? The logistic regression analysis presented in Table 4.2 confirms the importance of children’s age, but it also indicates:

• Children are significantly more likely to be included in the decision-making process when custody is settled privately rather than in court.

• In Quebec, children are much more likely to be consulted than elsewhere in Canada.

• Being consulted about living arrangements is related, statistically, to neither the gender of the child, nor that of the custodial parent.

Table 4.2 Impact of given variables on the probability that children in sole custody are consulted about living arrangements by separating parents, among children aged 4-15 years in 1998-99, whose parents separated in the previous two years, NLSCY, Cycle 3 (Logistic regression N = 731)

Variables Odds Ratio1 Child’s age at separation (continuous variable) 1.504*** Living arrangements • (With mother) 1.000 • With father 1.149 Type of custody agreement • (Court order) 1.000 • No court order 2.574*** Sex of child • (Boy) 1.000 • Girl 1.296 Region of residence in 1998-99 • (Canada, excluding Quebec) 1.000 • Quebec 3.512*** Reference category given in parentheses. 1 Coefficients significant at: ***p<.001; **p<.01; *p<.05; +p<.1.

New information on child support

Deciding on how to divide the financial responsibilities for children between parents who no longer share a household is a complex question and open to much dispute. The Federal Child Support Guidelines, introduced in 1997,12 were an attempt to establish more adequate and more consistent child support payments, and to reduce disagreement about support, by creating a more objective basis on which amounts would be estimated. The separations under observation in this analysis occurred precisely during the period following the introduction of these guidelines, and provide some insight into their initial effectiveness.

12 The same or similar guidelines were adopted in all provinces except Quebec, which designed its own formula. For a comparison, please refer to Chapter 3 of the publication, “Children Come First” (Department of Justice, 2002).

- 42 -

Reaching a child support agreement

As Figure 4.5 indicates, a child support agreement had been reached for 60% of children aged 4-15 years in 1998-99, whose parents had separated within the two preceding years. Almost three quarters of support agreements were reached privately, although the majority of these had been drawn up with legal assistance. Altogether, only 16% of children had support agreements contained in a court order, although there was one in progress for another 10%.

Not all separating couples reach an agreement on child support; among the children in this sample there was no child support agreement, and none in view, for three children in ten. Nonetheless, this figure is appreciably lower than that (38%) for separations occurring in the two years preceding Cycle 1 (Marcil-Gratton and Le Bourdais, 1999, Table 12). In other words, more couples appear to be reaching a child support agreement within a relatively short period after they separate. This result may support the claim that the 1997 Federal Child Support Guidelines have “reduced conflict and tension between parents, by making the calculation of child support orders more objective” and have “improved the efficiency of the legal process and most parents are now setting child support amounts without going to court” (Department of Justice, 2002, Volume 1, Minister’s preface).

Figure 4.5 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the existence and type of child support agreement, NLSCY, Cycle 3 (N=864)

Court-ordered agreement in progress

10%

No agreement30% Private formal

24%

Private informal 20%

Agreement60%

AGREEMENT: Court-ordered

16%

Why no child support agreement?

For the first time at Cycle 3, parents were asked why there was no support agreement. However, the fact that the reason fell outside the eight options in close to one-third of cases illustrates the complexity of the issue (Figure 4.6). Among the possible responses, the other parent’s inability to pay was the most common, given for 29% of children. Four other reasons were offered in

- 43 -

almost equal proportions (close to 10%): that child support had not been asked for; that parents could not agree; that other financial arrangements had been made; and that custody and cost-sharing were split between the parents. In a few cases, the custodial parent did not want child support or had been unable to locate the other parent.

Figure 4.6 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the reason why no agreement was made on child support payments, NLSCY, Cycle 3 (N=270)

29

4

10

3

7

7

8

32

0 5 10 15 20 25 30 35

Other parent unable to pay

Did not want child support

Did not ask for child support

Unable to locate parent

Parents cannot agree

Split-custody and cost sharing

Other financial arrangements

Other

%

Child support agreements and child support payments

Throughout Canada, child support is based on two principles: the joint financial responsibility of the parents and the sharing of responsibility in proportion to the resources available. Unless parents share custody time proportionally to their relative incomes, or unless the non-resident parent has insufficient income to make any contribution at all, it would be reasonable to assume that most child support agreements involve money exchanging hands. However, at Cycle 3, respondents who declared that a child support agreement had been reached were also asked whether payments were supposed to have been made for the child during the previous 12 months, or since the separation. Surprisingly, payments were expected for under two-thirds (63%) of the child support agreements reached for children, aged 4-15 years in 1998-99, whose parents separated at the end of the 1990s.

Why is it that over one-third of those with support agreements were not expecting child support payments to be made for the child? As no direct information is available at the NLSCY about the content of agreements, we do not know whether child support payments were actually stipulated in the agreement, or whether an alternative form of financial support, such as mortgage repayments, was agreed upon. It is possible, nonetheless, to glean some insight into the question. The figures in Table 4.3 represent the percentage of children for whom child support payments were expected among those for whom a child support agreement exists. This table shows:

- 44 -

• Payments were expected for four-fifths (81%) of children whose child support agreement was contained in a court order. This was so for under two-thirds (64%) of formal private agreements, and under half (47%) of informal private agreements.

• Whether or not support payments are anticipated depends largely on the child’s living arrangements after separation:

• With the mother: child support payments are most common when children are living all or most of the time with their mother after separation; child support was supposed to be received for almost three-quarters of those with a support agreement. On the other hand, support payments were not expected for a quarter of children in their mother’s custody—a proportion that reaches almost 40% when the agreement is a private informal one.

• With the father: payments are less common for children in their father’s custody; approximately one-third of those with an agreement expected child support payments to be made. This supports other studies showing that non-custodial mothers are rarely required to contribute to their children’s financial support. Seldom in paid employment before the separation, these mothers are unlikely to have the resources to make a contribution (Juby et al., forthcoming).

• Shared living arrangements: money rarely moves between households when children are in shared custody.13 Parents who share custody are often both in paid employment, and are more likely to each assume the costs of their child’s financial support when living in their household. It is also possible that the cooperative parenting required to manage shared custody arrangements may also lead to more flexible arrangements over children’s financial support.

• When child support agreements concern children of primary school age rather than younger or older children, child support payments are less frequently expected. This is also the case for agreements made in Quebec.

These factors are linked in various ways and we conducted a logistic regression analysis to understand the issue more clearly (see Table 4.4). This model confirms that, even controlling for the other variables, payments are expected significantly less frequently:

• In the case of private child support agreements, particularly if the agreement has not been formalized.

• If children are living with their father, or in shared custody, rather than with their mother after separation.

• When children are aged 6-11 years at separation, compared with younger children. This is also true for older children, though the coefficient is significant only at the 0.1 level.

13 Support payments are more common when children in shared custody are living mainly with their mother.

- 45 -

An interesting result relates to differences between Quebec and the rest of Canada. Table 4.3 showed that payments are expected in a lower proportion of child support agreements made in Quebec than in the rest of Canada. In the model, however, once other factors are controlled, the expectation of receiving payments is actually higher in Quebec than elsewhere in Canada. This suggests that the greater incidence of shared and father custody partly explains why the expectation of child support payments is lower in Quebec, though it may also be linked to differences in the way the Quebec maintenance enforcement program is implemented.

Table 4.3 Proportion of child support agreements that include child support payments, among children aged 4-15 years in 1998-99, whose parents separated in the two previous years, according to various characteristics, NLSCY, Cycle 3

Characteristics Proportion of child support

agreements that include child support payments

N 513 Proportion of child support agreements that include child support payments 63% Type of support agreement • Court order 81% • Private formal 64% • Private informal 47% Living arrangements • With mother, full-time: 74%

• Court order 84% • Private formal agreement 77% • Private informal agreement 62%

• With father, full-time 35% • Shared 14% Child’s age at separation (with mother) • 2-5 years 82% • 6-11 years 65% • 12-15 years 77% Region of residence in 1998-99 • Canada, excluding Quebec 66% • Quebec 57%

- 46 -

Table 4.4 Impact of given variables on the probability that a child support agreement does not include child support payments, among children aged 4-15 years in 1998-99, whose parents separated in the previous two years, NLSCY, Cycle 3 (Logistic regression odds ratios1 N = 518)

Variables Odds Ratio1 Type of support agreement • (Court order) 1.000 • Private formal 1.807* • Private informal 4.719*** Living arrangements • (With mother) 1.000 • With father 4.282** • Shared 23.564*** Child’s age at separation • (2-5 years) 1.000 • 6-11 years 2.808*** • 12-15 years 1.756+ Region of residence in 1998-99 • (Canada, excluding Quebec) 1.000 • Quebec .639+ Reference category given in parentheses. 1 Coefficients significant at: ***p<.001; **p<.01; *p<.05; +p<.1.

Are child support agreements adhered to?

The question about whether or not child support payments were part of the agreement, included for the first time at Cycle 3, has improved significantly the utility of subsequent questions referring to whether or not agreements are adhered to. In earlier cycles, it was assumed that a child support agreement entailed support payments, and all respondents declaring a child support agreement were asked about the regularity of payments. In cases where the agreement did not include child support payments, it is impossible to know how parents responded to this question. Respondents may have quite rightly declared that no payments had been received, thereby swelling the ranks of “non-paying” parents, and reinforcing, erroneously, the stereotype of parents neglecting their obligations towards children with whom they no longer live.

At Cycle 3, only in cases where child support was expected were parents asked about the regularity of payments. Overall, agreements that include the payment of child support were generally respected, at least within a relatively short period after the separation (Table 4.5). Payments were made regularly for the vast majority (84%) of children, and were mostly made on time (73%). Support was irregular for one child in eight (12%); and payments had never been received, or not for the last six months, for only one child in twenty-five.

- 47 -

Table 4.5 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the method and regularity of child support payments, and the proportion of payment received, NLSCY, Cycle 3

Characteristic % N 323 Regularity of payments • Regular and on time 73 • Regular but late sometimes 11 • Irregular 11 • No payments for at least 6 months 5 Total 100 Proportion of payment received • The whole amount 80 • Half the amount or more 6 • Under half 12 • None 2 Total 100 Although the question of the regularity of child support payments was included in all three cycles, in earlier cycles it was not possible to assess whether all, or only part of, the amount agreed upon was being paid. This information was collected for the first time in 1998-99, and showed that most child support was not only paid regularly, but was also paid in full. Over 80% of payments were for the total amount; at least half the sum agreed (and usually over three-quarters) was paid for an additional 6% of children. Less than half the amount was paid for one child in eight (12%) and no payment at all had been received for only 2% of the children for whom child support payments had been agreed upon. In addition, when payments were regular, they tended to be for the total amount: more than 90% of regular payments were paid in full, rarely the case for child support payments made at irregular intervals.

Cycle 1 data showed that child support payments agreed to privately were more often respected than those contained in a court order. This continues to be the case, as is evident in Figures 4.7 and 4.8. Private formal agreements are those in which support payments are both most regular and most often paid in full: child support is paid regularly for over 90% of children, usually on time, and the whole amount is paid in just under 90% of cases. The majority of child support contained in a court order is also paid regularly and in full, but the proportions drop to just over 70%. In other words, payments are intermittent and partial, or absent, for almost 30% of children whose financial support arrangements are included in a court order. Private informal agreements lie between these extremes, with child support less reliable than private formal agreements, but more so than settlements included in a court order.

- 48 -

Figure 4.7 Distribution of children aged 4-15 years in 1998-99, whose parents separated in the previous two years, according to the regularity of support payments and the type of support agreement, NLSCY, Cycle 3 (N=348)

7973

66

1414

6

79

21

5 60

0%

20%

40%

60%

80%

100%

Private formal agreement Private informal agreement Court ordered agreement

Type of support agreement

None for 6+months

Irregular

Regular, late at times

Regular, on time

Figure 4.8 Distribution of children aged 4-15 years in 1998-99, whose parents separated

in the previous two years, according to the proportion of child support payments made and the type of support agreement, NLSCY, Cycle 3 (N=349)

8878 72

5

66

710 19

53

0%

20%

40%

60%

80%

100%

Private formal agreement Private informal agreement Court ordered agreement

Type of support agreement

NoneLess than halfMore than halfWhole amount

- 49 -

DISCUSSION

The last decades of the 20th century were marked by profound changes in the family: declining fertility levels, increasing divorce, and the decline of marriage’s monopoly, first for the formation of conjugal unions, and more recently, as the context for establishing a family. The impact of these shifts on Canadian children is equally profound, as changes in their parents’ conjugal and parental life translate into family transitions in their own. This second phase of research into the impact of parents’ transitions on children’s family environment and economic well-being has focused on what is, for most children, the first, and perhaps most significant, change during childhood—their parents’ separation. More specifically, this research involved: a) updating information on the context within which children are born and its impact on their subsequent family life course, with particular reference to parental separation; b) exploring the arrangements made by separated parents concerning their children’s care after separation, with the focus on shared custody; and c) analysing new information on the custody and child support arrangements made by separated parents.

Children are born into a variety of contexts because they arrive in their mother’s and father’s life at different points. The birth of a first child to a young single mother, for example, may be the first “event” in the mother’s family life course. Most children arrive further on, to parents who have made particular lifestyle choices—to marry before starting a family, for example—or who are already on a second or third family. The point at which children arrive in their mother’s and father’s life course is linked to the way in which their own family life during childhood is likely to evolve, although the nature of this link is not clear. Children born in a second family, for example, are more likely to experience their parents’ separation than children born into their parents’ first family, as are children born within a common-law union compared with those born within marriage.

However, these factors are themselves in evolution. The context at birth changed dramatically during the last two decades, and almost one-third of Canadian children were born outside legal marriage by the turn of the century. Moreover, while the popularity of cohabitation in Quebec largely accounts for the rise in common-law union births in Canada, there has been a significant increase in the phenomenon in all regions. Compared with children born in the early 1980s, for example, those born at the end of the 1990s were twice as likely to be born to a cohabiting couple in Ontario and the Prairies, and almost three times as likely in Eastern Canada. Similarly, the proportion of children born within the second family established by their mother or father rose from 11% among the oldest cohorts to 18% among the youngest.

Overall, the probability of parental separation rose rapidly throughout Canada for children born during the 1980s, and then leveled off by the early 1990s, leaving separation levels high, but no longer increasing. However, a rise, during the 1990s, in the proportion of children born outside a union means that this second entry point into single-parent family life may assume a greater relative importance in the next decade, particularly in regions where separation levels are lower than average and out-of-union births higher. Already, regions of Canada with low levels of separation find themselves with as high a proportion of children experiencing life with a lone parent as other regions. The Atlantic provinces, for example, is one of the only regions in which children born in a union have a three in four chance of still living with both parents at their tenth

- 50 -

birthday. However, the high proportion of births to single mothers, with one in six (16%) babies born outside a union, raises the likelihood of a child experiencing life with a lone parent to levels similar to those in the other provinces.

This evolution is all the more important because the pathway into a lone-parent family is strongly linked to the type of lifestyle a child is likely to have during the episode. Children in lone mother families created after separation or divorce, for instance, enjoy a higher standard of living, on average, than children born to a young single mother (Peron et al. 1999). This has obvious social policy implications, with a higher proportion of lone-parent families in need of financial assistance in some provinces than in others.

With no father present, many single mothers have to assume full responsibility for their baby’s physical and economic support. When couples separate, on the other hand, these responsibilities have to be renegotiated, with arrangements generally recorded in the form of custody and child support agreements. One problem, raised earlier, is that the published divorce statistics do not make it possible to follow the evolution of children’s living arrangements following separation. Recently, in the Child-Centred Family Justice Policy, the Department of Justice suggested replacing the terms “custody” and “access” with a new model based on parental responsibilities, with both parents responsible for their children’s well-being after separation or divorce. Arrangements made by parents at separation would include a residential schedule setting out the time spent by children with each parent, on the one hand, and decision-making responsibilities related to children’s health, education and other matters, on the other. Clarifying the different elements of children’s care after separation, and recording decisions related to a child’s physical as well as legal custody, could go a long way towards improving current information on the way custody and living arrangements are evolving.

Sharing physical custody at separation is growing in popularity, despite certain obvious drawbacks of this type of arrangement. Even though it seems to be a difficult arrangement to maintain over time and often transforms into sole custody, sharing custody for a period after separation appears to lay the basis for children’s continued long-term involvement with both parents. It may be an important step in the transition towards a secure relationship with a non-resident parent, giving parents and children alike the time to develop new and secure foundations for their relationship. However, not only does this arrangement require a high degree of cooperation between separated parents, the additional cost in having two family homes puts it beyond the means of many, and the need for geographic proximity makes it unworkable for others. With research in the United States suggesting that non-resident parents remain more involved with their children if “legal” custody is shared (Seltzer 1991, 1998), in cases where parents are unable to share the residential responsibility for their children, each parent having responsibility over other aspects of the child’s life could have a positive effect.

Although problems with the questionnaire limited the scope of the analyses, the new questions about custody and child support, added at Cycle 3, provide some important insights into the separation process at the end of the 1990s. One issue currently under debate is whether children should be given more of a voice in decisions made about their future at divorce proceedings. Courts are wary of exposing children to conflict between parents, or of asking children to “choose” between their parents. Nevertheless, the analysis in this report shows that the majority of children over the age of seven or eight years are consulted by their separating parents,

- 51 -

particularly when custody is settled privately rather than through the legal system. In other words, children are being asked to voice their preferences; parents consider that children have a right to their say, and the opinion they express is taken seriously. It could be argued that not giving the same rights to children whose parents are unable to reach an agreement out of court is simply penalizing them further.

The information on child support payments improved considerably with the addition of questions concerning a) whether having a child support agreement meant that child support payments were expected and b) the proportion of the payments received. The analysis indicates that the vast majority of payments are made both regularly and on time, at least during the relatively short period after separation under observation. The problem seems above all to be that of negotiating an agreement, rather than its implementation once it has been reached. However, the fact that more separating couples are reaching a child support agreement, and more rapidly, suggests that the Child Support Guidelines introduced in 1997 have had some success in helping separating couples share financial responsibilities for their children. Nonetheless, there is still some way to go; perhaps the most pressing issue is related to the prohibitive financial costs involved when parents are unable to agree and which may ultimately compromise the best interests of the child.

Finally, a theme that has reappeared at several points during the text is the distinctive nature of Quebec, not only with regard to common-law unions but also in relation to the way parental responsibilities are shared at separation. In Quebec, shared living arrangements and sole father custody are more common; shared custody is more durable; agreements made about custody and access are generally more strictly adhered to; and children are more often consulted about these arrangements. Why is this? Is it a result of the different approach to separation and divorce in Quebec law (see Department of Justice, 2002)? Or is it a social rather than a legal phenomenon? Are family roles less defined along gender lines in Quebec? Is the symmetry in the discourse and practice of parental roles, identified in a study of Montreal couples with shared physical custody (Côté, 2000), more prevalent in Quebec than elsewhere in Canada? This is an interesting avenue for further research, not least to discover whether the differences between Quebec and the rest of Canada extend to the impact of separation on children.

- 53 -

REFERENCES

Bessner, R. 2002. The Voice Of The Child In Divorce, Custody And Access Proceedings. Background report commissioned by the Family, Children and Youth Section, Department of Justice Canada. October 2002. Cat: 2002-FCY-1E.

Côté, D., 2000. La garde partagée. L’équité en question. Les Éditions du remue-ménage, Montréal, pp. 202.

Department of Justice. 2002. Children Come First: A Report to Parliament Reviewing the Provisions and Operation of the Federal Child Support Guidelines. vols. 1 and 2. Cat: J2-186/2002-1E and 2E.

Institut de la Statistique du Québec. 2000. Table entitled: Répartition des naissances selon l’état matrimonial de la mère par région administrative, Québec, 1998, on the ISQ website: www.stat.gouv.qc.ca.

Juby, H., N. Marcil-Gratton and C. Le Bourdais. 2001. “A Step Further in Family Life: The Emergence of the Blended Family,” Report on the Demographic Situation in Canada—2000. Ottawa: Statistics Canada. Cat: 91-209.

Juby, H., C. Le Bourdais and N. Marcil-Gratton. 2003. Linking Family Change, Parents’ Employment and Income, and Children’s Economic Well-Being: A Longitudinal Perspective. Ottawa: Child Support Team, Department of Justice Canada Research Report (forthcoming).

Le Bourdais, C., G. Neill and N. Marcil-Gratton. 2000. “L’effet du type d’union sur la stabilité des familles dites ‘intactes’,” Recherches sociographiques. XLI, 1: 53-74.

Marcil-Gratton, N. 1998. Growing up with Mom and Dad? The intricate family life courses of Canadian children. Ottawa: Statistics Canada, Cat: 89-566-XIE.

Marcil-Gratton, N. and C. Le Bourdais. 1999. Custody, Access and Child Support: Findings from the National Longitudinal Survey of Children and Youth. Ottawa: Child Support Team, Department of Justice Canada Research Report CSR-1999-3E.

Péron, Y., H. Desrosiers, H. Juby, C. Le Bourdais, E. Lapierre-Adamcyk, N. Marcil-Gratton and J. Mongeau. 1999. Canadian Families at the Approach of the Year 2000. Ottawa: Statistics Canada, 1991 Census Monograph series, Cat: 96-321-MPE9804.

Seltzer, J.A.. 1991. “Relationships Between Fathers and Children Who Live Apart: The Father’s Role after Separation.” Journal of Marriage and the Family. 53:79-101.

Seltzer, J.A.. 1998. “Father by Law: Effects of Joint Legal Custody on Nonresident Fathers’ Involvement with Children.” Demography, 35, 2: 135-146.

- 54 -

Statistics Canada. 1990. Health Reports. Ottawa: Minister of Supply and Services, Canada. Cat: 82-003: vol. 2, no. 1.

Statistics Canada. 1992. Health Reports: Divorce 1990. Ottawa: Minister of Industry, Science and Technology, Canada. Cat: 82-003S17: vol. 3, no. 4.

Statistics Canada.1997. Divorces 1995. Ottawa: Minister of Industry. Cat: 84-213-XMB.

Statistics Canada. 1997. Divorces 1999 and 2000. Ottawa: Minister of Industry. Cat: 84F0213-XPB.